teicoplanin aglycone core-shell chiral stationary phases based on teicoplanin and Enantioseparation of ß -amino acids by liquid chromatography using Journal of Chromatography A

(1)

ContentslistsavailableatScienceDirect

Journal of Chromatography A

journalhomepage:www.elsevier.com/locate/chroma

Enantioseparation of ß ² -amino acids by liquid chromatography using core-shell chiral stationary phases based on teicoplanin and

teicoplanin aglycone

Dániel Tanács

^a

, Róbert Berkecz

^a

, Aleksandra Misicka

^b

, Dagmara Tymecka

^b

, Ferenc Fülöp

^c

, Daniel W. Armstrong

^d

, István Ilisz

^a^,^∗

, Antal Péter

^a

aInstitute of Pharmaceutical Analysis, Interdisciplinary Excellence Centre, University of Szeged, Somogyi B. u. 4, H-6720 Szeged, Hungary

bDepartment of Chemistry, University of Warsaw, Pasteura str. 1, 02-093 Warsaw, Poland

cInstitute of Pharmaceutical Chemistry, University of Szeged, Eötvös utca 6, H-6720 Szeged, Hungary

dDepartment of Chemistry and Biochemistry, University of Texas at Arlington, Arlington, TX 76019-0065, USA

a rt i c l e i nf o

Article history:

Received 30 April 2021 Revised 18 June 2021 Accepted 28 June 2021 Available online 5 July 2021 Keywords:

ß²-amino acids liquid chromatography

macrocyclic glycopeptide-based chiral stationary phases

kinetic and thermodynamic characterization, core-shell particles

a b s t r a c t

Enantioseparation of nineteenß²-amino acids has been performed byliquid chromatographyon chiral stationary phases based on native teicoplanin and teicoplaninaglycone covalently bonded to 2.7 μm superﬁciallyporous silica particles. Separations werecarried out inunbuffered (water/methanol), buffered[aqueoustriethylammoniumacetate(TEAA)/methanol]reversed-phase(RP)mode,andinpolar- ionic(TEAAcontainingacetonitrile/methanol)mobilephases.EffectsofpHintheRPmode,acidandsalt additives,aswellascounter-ionconcentrationsonchromatographic parametershavebeenstudied.The structureofselectands(ß²-aminoacidspossessingaliphaticoraromaticsidechains)andselectors(native teicoplaninorteicoplaninaglycone)wasfound tohave aconsiderableinﬂuenceonseparation performance.AnalysisofvanDeemterplotsanddeterminationofthermodynamicparameterswereperformed tofurtherexploredetailsoftheseparationperformance.

ThisisanopenaccessarticleundertheCCBYlicense(http://creativecommons.org/licenses/by/4.0/)

1. Introduction

Inthepastdecade,considerableinteresthasbeenpaidtopep- tides containing ß-amino acids (ß-peptides). With an additional carbon atom betweenthe amino andcarboxylic groups, theseß- amino acids can adopt various stable secondary structures with furtherfunctionalizationpossibilitiesenhancingthenumberofpo- tential applications[1].Unliketheir analogs,theseß-aminoacids arenotreadilysusceptibletohydrolysisorenzymaticdegradation.

Consequently, peptides with incorporated ß-amino acids exhibit enhanced stability [2]. Additionally, chimeric peptides (mixed

α

^-

andß-peptides)consistingof

α

^-ând^ß-aminoâcidsâreôf^consid-

erable interestaspeptidomimeticsinan increasingrangeofther- apeutic applications [3,4]. Depending on the position ofthe side chainonthe3-aminoalkanoicacidskeletonß²-andß³-aminoacids can be differentiated. The syntheses of ß²-amino acids in enan- tiomerically pure formaremuch morechallengingthan their ß³-

∗Corresponding author: István Ilisz, Institute of Pharmaceutical Analysis, Univer- sity of Szeged, Somogyi B. u. 4, H-6720 Szeged, Hungary

E-mail address: ilisz.istvan@szte.hu (I. Ilisz).

analogs[5].Thesynthesis ofß²-amino acidsinracemicformand their subsequent enantioseparation currently is the most practi- cal and effectiveapproach to obtain enantiopure ß²-amino acids.

Chromatographic datarelatedtothe separationandidentiﬁcation of

β

³^-amino ^acid enantiomers have been reported in the literature[6-8].However,relativelylittleinformationisavailableonthe separation of ß²-amino acid enantiomers. The enantioseparation ofa fewß²-aminoacids haverecentlybeencarriedout by direct high-performanceliquid chromatography(HPLC)methods onchi- ral stationary phases(CSPs) based on (+)-(18-crown-6)-2,3,11,12- tetracarboxylic acid [9,10], macrocyclic glycopeptides [11,12], and Cinchonaalkaloids[13,14].

Core-shell particles (superﬁcially porous particles, SPPs) and sub-2 μm fully porous particles (FPPs) are expected to provide high-throughput and effective separations of a variety of chiral molecules in ultra-high-performanceliquid chromatography (UH- PLC). Teicoplanin, teicoplaninaglycone, vancomycin, or isopropyl- cyclofructancovalentlybondedtosub-2 μmor2.7μmSPPswere successfully applied for the enantioseparation of native and N- protected

α

^-amino âcids ând ^small ^peptides ûnder ^LC ^[15-18]^,

andsupercriticalﬂuidchromatography(SFC)conditions[19,20].Te-

https://doi.org/10.1016/j.chroma.2021.462383

(2)

Figure 1. Structure of ß²-amino acids

icoplanin and teicoplaninaglycone covalently attachedto 1.9μm FPP silica gel with narrow size distribution exhibited excellent separation performances for native proteinogenic amino acids in both LC and SFCmodalities [21].The newsynthetic route developed for bonding teicoplanin to sub-2 μm or 2.7 μm SPPs and to sub-2μm FPPsendowedtheselector withazwitterionicchar- acter [22,23]. Ion-exchange-type CSPs are also being developed for UHPLC purposes. For example, tert-butylcarbamoyl(8 S,9 R)- quinine wascovalently bondedto 1.9μm [22] or to 2.7μm [24- 30] SPPs, and to 3.0 μm and 1.7μm FPPs [28]. Lämmerhofer et al. [30] in chiral × chiral two-dimensional UHPLC applied tert- butylcarbamoyl(8 S,9R)-quinine andtert-butylcarbamoyl(8 R,9S)- quinidine selectors bondedto 2.7μm SPPs forthe separation of enantiomers of native proteinogenic

α

^-âmino âcids âfter ^peptide

hydrolysis. A survey of literature data revealedthat enantiosepa- rationsunderUHPLCconditionswereperformedforproteinogenic

α

^-âminoâcids ^with^theônlyêxceptions^being^the enantioseparationof

γ

-aminobutyricacid[27]andß-Ala[28,34].

The present studyprovides results for the utilization of CSPs based on macrocyclic glycopeptides covalently bondedto 2.7μm SPPs forthe enantioseparationof19unusual ß²-aminoacids.Ex- perimentswereperformedutilizingcolumnswith3.0and2.1mm internaldiameter(i.d.)basedonteicoplanin-andteicoplaninaglycone. RP and polar-ionic mobile phases were applied in separations. Effectsofthenatureandconcentrationofthemobilephase components,acidandsaltadditivesundervariousconditions,and pHinreversed-phase(RP)modewerestudied.Togaindetailedin- formationaboutthechiralrecognitionprocess,structure–retention (selectivity) relationships were evaluated by taking into account thestructuralfeaturesofboththeanalytesandselectors.Analysis of van Deemterplots served asa basis forthe kinetic investiga- tions,whilethetemperaturedependencestudyallowedthermody- namiccharacterization.Inafewcases,elutionsequencesalsowere determined.

2. Experimental

2.1. Chemicalsandmaterials

Nineteen racemic amino acids (1-19) were synthesizedasde- scribed earlier[13].(Forstructures seeFig.1). Enantiomers(R)-2,

(S)-5and(S)-6weregenerousgiftsfromProf.D.Tourwé (VrijeUni- versiteitBrussels,Brussels,Belgium).

Methanol (MeOH), acetonitrile (MeCN), and water of LC-MS grade, NH₃ dissolved in MeOH, triethylamine (TEA), formic acid (FA), glacial acetic acid (AcOH), ammonium formate (HCO₂NH₄), andammoniumacetate(NH₄OAc)ofanalyticalreagentgradewere from VWR International (Radnor, PA, USA). The pH reported for thereversed-phasemobilephase istheapparent pH(pH_a),which was adjusted directly in the aqueous triethylammonium acetate (TEAA)/MeOHsolutionswiththeadditionofAcOH.

2.2. Apparatusandchromatography

LC measurements were carried out on a Waters® ACQUITY UPLC® H-Class PLUS System with Empower 3 software (Waters) and components asfollows: quaternary solvent manager, sample manager FTN-H, column manager, PDA detector, and QDa mass spectrometer detector.Theparameters forthe QDadetectorwere setasfollows:positiveionmode,probetemperature, 600°C,cap- illaryvoltage,1.5V,conevoltage,20V.

Chiral columns, based on teicoplanin (TeicoShell, T) and teicoplaninaglycone (TagShell, TAG) attachedcovalentlyto thesur- face of 2.7 μm SPPs, were applied in this study. The core di- ameter and shell thickness of the SPPs were 1.7 μm and 0.5 μm,respectively.Allcolumns(AZYP,LLC,Arlington,TX,USA)have 100×3.0mmi.d.or100×2.1mmi.d.dimensions(abbreviations forcolumns:T-3.0andT-2.1;Tag-3.0andTag-2.1).

Stock solutions of analytes (1-10 mg ml^–1) were prepared in MeOH anddiluted with themobile phase. The dead-time (t₀) of thecolumnswasdeterminedby0.1%AcOHdissolvedinMeOHand detectedat210or256nm.Inallexperimentsaﬂowrateof0.3ml min^–1providedeﬃciencyandfastanalysisforbothcolumndimen- sions,whilethecolumntemperaturewassetat20°C (ifnot oth- erwisestated).

3. Resultsanddiscussion

Basedon theirside chain,theinvestigatedß²-aminoacidscan be divided into two sub-groups: those that contain an aliphatic moiety or an aromatic moiety. The branch or the length of the aliphaticmoietyorthenatureandpositionofsubstituentsonthe

(3)

aromaticringinﬂuencesthesizeandpolarityofthemoleculeand isexpectedtohaveaconsiderableeffectonselector–analyteinter- actions.

3.1. EffectofpHonretentionandseparationperformance

ThepKvalueofcarboxylicgroupsofteicoplaninandteicoplanin aglycone (playing important role in the retention mechanism) is approximately2.5.ThepKvaluesoftheaminogroupsofß²-amino acids 1-19 are in the range 10.16–10.32. The corresponding values for the carboxylic moieties of 1-12, 19 are between 4.10- 4.50, whereas for 13-18 they are between 3.20-3.60 (calculated with MarvinSketch v. 17.28 software,ChemAxon Ltd., Budapest).

Therefore, the charge of macrocyclic glycopeptide-based stationary phasesandanalytesis sensitiveto pHandalterstheinterac- tions betweentheCSP andtheanalyte.Toreveal thepossibleef- fects of pHa on retention, selectivity, andresolution of ß²-amino acids bearing aliphatic (3) and aromatic side chain (9) were selected andtheirretentionbehavior wasinvestigatedon T-3.0and TAG-3.0 columns in the RP mode applying eluents consisting of aq.TEAA/MeOH (90/10 v/v and 30/70 v/v) with a constant TEAA concentration of20.0mM) andvarying pHa of themobile phase betweenpHa3.5−6.5. Underthestudiedconditions,theretention factorsoftheﬁrstelutingenantiomer(k₁)usuallychangedslightly withincreasingpHaforbothanalytes,andonlyanalyte9exhibited moderate increasesin the aq.TEAA/MeOH 30/70 (v/v) eluent (Fig.

S1). Interestingly,

α

ând ^RS increased more significantly in both mobile phase systems, especially foranalyte 3, with the highest valuesobtainedabovepHa5.0(Fig.S1).Based ontheirpKvalues, teicoplanin, teicoplaninaglycone, andß²-aminoacids are present inionizedformunderthesemobile phaseconditions.Thatis,the observed behavior isprobablydue toincreasedionic interactions betweentheprotonatedaminogroupoftheanalyteandthedepro- tonatedcarboxylicgroupoftheselector.Theionicstructuresaffect either directly the Columbic or dipolar interactions between the analyteandselector,orinfluenceindirectlytheretentionbychang- ing theconformation ofthe macrocyclic glycopeptides. Toobtain thehighestresolutionandselectivityaneluentpHaof5.0orabove canberecommendedfortheenantioseparationofß-aminoacids.

3.2. Effectsofmobilephasecompositiononthechromatographic performance

Employing analytes 3 and9, first, the effects of five different mobile phaseadditives(salts oracids,namelyHCO₂NH₄,NH₄OAc, TEAA, FA,and AcOH) were studied on the chromatographic per- formanceofT-3.0andTAG-3.0CSPs.Experimentswereperformed withaconstantaqueoustoorganicsolventratio(H₂O/MeOH90/10 v/v)andaconstantadditiveconcentration(20.0mM,calculatedfor thewholeeluentsystem).Inthecaseoforganicsalts,thepHawas adjusted to5.0by theadditionofthe correspondingacid. Mobile phasescontainingsolely20.0mMFAorAcOH(withoutpHadjust- ment) resultedinunresolvedpeakswithratherpoorpeak shapes (Fig.S2). Incontrast,whenHCO₂NH₄,NH₄OAcorTEAAwereused asmobilephaseadditive,resolutioncouldbeobtained.Employing TEAA has ledto symmetricalpeak shapes, very goodefficiencies, andselectivities.Therefore,infurther experiments,TEAAwasthe favoredmobilephaseadditive.Itisworthmentioningthatregard- ing MS-baseddetection,NH₄OAcoffershighersensitivitywithac- ceptablepeakshapesandresolution.

MeOH and MeCN organic modiﬁers are used commonly in amino acid separations [9]. The nature and concentration of the mobile phase components can exert considerable effects on retention and separation. Therefore, we next investigated the effects oforganic modiﬁerson the enantioseparationof analytes3

and9utilizingT-3.0 andTAG-3.0CSPs. Figure 2showsthe chromatographic ﬁgures of merit for the separations of analytes 3 and 9 in three different eluent systems. In unbuffered RP mode (a), mobile phase compositions of H₂O/MeOH 90/10–10/90 (v/v), in buffered RP mode (b), aq.TEAA/MeOH 90/10-30/70 (v/v) containing 20 mM TEAA and pHa 5.0, and in polar-ionic mode (c), MeCN/MeOH90/10–10/90(v/v) containing20mM TEAAwere applied.

Intheunbufferedeluentsystem(Fig.2A),k₁increaseswithin- creasingMeOHcontent,however,not inthewholerangestudied.

Theobservedphenomenonisatleastpartlyforthelowersolubil- ity of amino acids withpolar character in the lesspolar MeOH.

Theobservedminimuminthecurveforanalyte9indicatesanin- creasedhydrophobic interaction at higherwatercontent. Regard- ing

α

^and^RS values,theyincreasewithincreasingMeOHcontent.

Interestingly, comparing the two CSPs, k₁ values were higher on theT-3.0column,whilehigher

α

^and^RS valueswereregisteredon TAG-3.0.,whichmayindicate reducednonselectiveinteractions in thelattercase.

Under buffered RP conditions (Fig. 2B), similar to the un- bufferedeluents,aslightincreaseink₁,

α

^,^and^RS valueswasreg- isteredwithincreasingMeOH content.Asan exception,analyte 9 on theTAG-3.0 columnﬁrst showeda moderateincrease, then a slightdecreasefork₁.Comparingthesetwoeluentsystems, are- markable difference between chromatographic performances can benoted.InthepresenceofTEAA,higher

α

^and^RS valuesareob- tainedwithsigniﬁcantlylowerretentions,suggestingapronounced suppressionofnonselectiveinteractions betweentheanalytesand thestationaryphasebythesaltaddition.

TheresultsobtainedwiththeapplicationofmixturesofMeCN andMeOH along with acid andbase additivesin the polar-ionic mode aredepictedin Fig.2C. Withvariation inthecomposition of the eluent the acid–base equilibrium and proton activity will be changed. At high MeCNcontent, the solvation of polar amino acidsintheaproticsolventdecreasesresultinginhighretentions, whilethe increasing ratioof protic MeOHfavors thesolvation of polaramino acids, i.e.,retention decreases. The change of

α

^and

R_S values exhibited trends similar to those discussed above. The improvedselectivity with increasing MeOHcontent suggeststhat hydrogenbondingmaynot playamajorroleintheseenantiosep- arations.

3.3. Effectsofthecounter-ionconcentration

The stoichiometric displacement model [31] is applied fre- quently to describe the retention behavior based on ion-pairing andion-exchangemechanisms,predictingalinearrelationshipbe- tween thelogarithm ofthe retentionfactor andthelogarithm of thecounter-ionconcentration,

logk= logKZ−Zlogccounter−ion (1)

whereZ isthe ratio ofthe numberof chargesof the cationand thecounter-ion, whileKz describestheion-exchangeequilibrium.

If an ion-exchange mechanism exists, plotting log k against log ccounter-ion will result in a straight line witha slope proportional totheeffectivechargeduringtheion-exchangeprocess,whilethe interceptprovidesinformationabouttheequilibriumconstant.

Toprobethepotencyofthesimpledisplacementmodelinour case,experimentswerecarriedout onT-3.0 andTAG-3.0CSPsap- plying mobilephasesb,aq.TEAA/MeOH (90/10v/v, pHa ≈5.5)and c,MeCN/MeOH(10/90v/v) both containing5.0-160 mMTEAA. In a cation exchange process in the presence of TEAA, the proto- natedtriethylammoniumionactsasacompetitor.Theresultspre- sented in Fig. 3, deﬁnitely indicate the applicability of the stoichiometric displacement model, i.e., they support the involve- ment ofion-interactionprocesses intheretention mechanism. In

(4)

Figure 2. Effect of bulk solvent composition on chromatographic parameters for analyte 3 and 9 applying different mobile phase systems

Chromatographic conditions: column, T-3.0 and TAG-3.0; mobile phase, a, H 2O/MeOH (90/10–10/90 v/v ), b , aq. TEAA/MeOH (90/10-30/70 v/v ), concentration of TEAA in mobile phase 20.0 mM and the actual pH of the mobile phase, pH a5.0, c , MeOH/MeCN (90/10–10/90 v/v ), concentration of TEAA in mobile phase, 20.0 mM; ﬂow rate, 0.3 ml min ^–1; detection, 210-258 nm; temperature, 20 °C; symbols, on T-3.0 for analyte 3 , , for analyte 9 , █, on TAG-3.0 for analyte 3 , , for analyte 9 ,

Figure 3. Effect of ion content on retention factor of the ﬁrst eluting enantiomer, k 1 for analytes 3 and 9 Chromatographic conditions: column, T-3.0 and TAG-3.0;

mobile phase, A, aq. TEAA/MeOH (90/10 v/v ), concentration of TEAA in mobile phase, 5.0-160 mM, B, MeCN/MeOH (10/90 v/v ), concentration of TEAA in mobile phase, 5.0-160 mM; ﬂow rate, 0.3 ml min ^–1; detection, 210-258 nm; temperature, 20 °C;

symbols, on T-3.0 for analyte 3 , , for analyte 9 , █, on TAG-3.0 for analyte 3 , , for analyte 9 ,

thisstudy,linearrelationshipswere foundbetweenlogk₁ vs. log ccounter-ion,withslopesvarying betweenabout(–0.10)and(–0.23).

Inan earlierstudy,slopes around–1.0 werefoundforstrongion- exchangers, where the selector and selectand act in almost fully ionizedform [32].In absoluteterms,thesmallerslopes observed reveal a marked difference between the strong and weak ion- exchanger-basedCSPs[33].Inthecaseofweakion-exchangerCSPs, theretention(affectedbythepHandtheionicstateoftheselector andanalyte)canbereducedwiththeenhancementofthecounter- ionconcentration,butonlytoalimitedrange.Itisworthmention- ing thatonbothCSPs,practicallyequalslopes werecalculatedfor each enantiomer, i.e., no signiﬁcant difference could be observed

inthe enantioselectivitieswithvarying counter-ion concentration (datanotshown).

3.4. Effectsofstructuresofselectorandanalyteonretentionand selectivity

The structure of both the chiral selector and the analyte af- fects considerably their interactions resulting in different retention and separation characteristics. To gain a set of chromatographic data, screening of the enantioseparation of 19 ß²-amino acidswasperformedonfourteicoplaninandteicoplaninaglycone- basedcolumnsinthreedifferentmobilephasesystems:unbuffered RP(a,H₂O/MeOH30/70v/v),bufferedRP(b,aq.TEAA/MeOH30/70 v/v, containing 2.5 mM TEA and 5.0 mM AcOH, pH_a 5.5), and a polar-ionicmobile phase (c,MeCN/MeOH30/70 v/v,containing 2.5mMTEAand5.0mMAcOH).Therelatedchromatographicdata aresummarizedinTablesS1−S4.Allstudiedß²-aminoacidswere baseline-separatedon atleastone CSP,andoften withbothCSPs withinthreetoﬁveminutesdependingonthenatureofanalytes, mobilephase,andinnerdiameterofcolumns.The overallsuccess rate ofthe enantioseparations is depicted in Fig. S3. Taking into account the time needed for the analyses, application of mobile phase a and b seemed to be more favorable (Tables S1−S4). It should benoted, that the analysistime obtainedhereisthree to tentimeslowerthan thatobserved earlieron 5μmparticlesand 4.6 mm i.d. columns [10-13]. It was also observed that, in most cases,ß²-aminoacidspossessingaliphaticsidechains(analytes1- 8)exhibitedslightlysmallerR_Svaluesthananalyteswitharomatic sidechains(9-19).Thisisinspiteoftheir similarenantioselectiv- ity(1.30<

α

<2.20).Foranalytes9-19, inalmost all cases,RS >

1.5wasobtainedonallfourcolumnsappliedwithanyofthethree mobilephasesystems(exceptionswerecompounds12and13).

(5)

Figure 4. Dependence of retention factors of the first eluting enantiomer ( k 1) and separation factors ( α) of analytes 1-6 on the Meyer substituent parameter ( V â) Chromato- graphic conditions, column, T-3.0, T-2.1, TAG-3.0 and TAG-2.1; mobile phase, A , aq. TEAA/MeOH (30/70 v/v ), concentration of TEA and AcOH in mobile phase 2.5 and 5.0 mM, respectively and the actual pH of the mobile phase, pH a5.5, B, MeCN/MeOH (30/70 v/v ), concentration of TEA and AcOH in mobile phase 2.5 and 5.0 mM, respectively; flow rate, 0.3 ml min ^–1; detection, 210-258 nm; temperature, 20 °C; symbols, for T-3.0 █, for TAG-3.0 , for T-2.1 and for TAG-2.1

Inordertodeterminethespeciﬁcstructuraleffectsofanalytes possessing alkyl side chains on chromatographic data such ask₁ and

α

^,^theêffectôf^the^volumeôf^theâlkyl substituents(Vâ)was investigated.Thestericeffectofasubstituentonthereactionrate can be characterized by the size descriptor of the molecule, Vâ [34].TheVâ valuesforMe,Et,Pr,Bu,2-Pr,and2-Bumoietiesare 2.84,4.31,4.78,4.79,5.74,and6.21×10⁻²nm³,respectively.Note, thattherearenoVâvaluesavailablefor6-methylheptanoic(7)and 5-cyclohexylpentanoic (8) moieties.Valuesofk1 and

α

^showed^a

good correlation withVâ on all studiedcolumns inall threeelu- entsystems.AsthedatapresentedinFig.4confirmthevolumeof thealkylsubstituentsmarkedlyinfluencedk₁:abulkiersubstituent hindered the interactions betweenthe selector andanalyte lead- ing to reducedretention. Since the difference in the interactions of thetwo enantiomeric analyteswiththe CSPdiffered considerably,anenhancedchiralrecognitionwithhigherVâvaluescouldbe observed. Itshould benotedhere, thatnot onlythe positionand bulkiness ofthe substituent butalso the steric effect may heav- ilyinfluenceretentionbehaviorandchiralrecognitionofß²-amino acids.

Comparingtheseparationofanalytes9-19possessingaromatic orsubstituted aromaticside chainsto analytes1-8.showshigher R_S values for analytes 9 – 19. In most cases, the R_S was above 1.5andonlyanalytes12and13exhibitedpoorerresolution(Table S1−S4). The mostrelevantandoptimized dataofseparations are depictedinTable1.Thepresenceofanaromaticmoietyinsteadof analiphaticsidechainin9-19probablyimproves

π

^–

π

-interactions between theenantiomers andthe chiralselector and contributes to betterchiral recognition.Enantiomersofanalyte 12possessing an additional 4-dimethylamino moiety (pKa 5.0, calculated with Marvin Sketch v. 17.28 software, ChemAxon Ltd., Budapest) were baseline-separatedonlyinmobilephasesbandc,wheretheionic strengthcouldbekeptataconstantlevel.

Analytes 11,13,and14possessamethyl,chlorine,orhydroxyl substituent at position 4 of the aromatic ring, giving

π

^-^basic ^or

π

^-âcidic ^character ^to ^the ^molecules. ^Figures ⁵ ând ^S4 âre ^chro-

matogramsthat illustratetheseparationperformance obtainedon TAG-3.0andT-3.0CSPsintwodifferenteluentsystems.Themethyl andchlorine moietiesshow slighteffectsonretention,selectivity, andresolution,whilethehydroxylmoietiesandtheirpositions in analytes14vs.15 affectconsiderablytheseparationperformance.

The3-positionofthehydroxylmoietyprobablyfavorsstericinter- actions betweenselector andanalyte resultingin higherselectiv- ity andresolution,in particular,on the TAG-3CSP in H₂O/MeOH (30/70v/v)mobilephase(Fig.5A).Thesedifferences,especiallyin resolution,canbeobservedinFig.S4AandS4B.

Analytes16-18possessanadditionaletherO-atom,whichisca- pable of H-bond interactions, while 19 bears a naphthyl moiety,

whichmayfacilitatestronger

π

^–

π

-interactions.Allthesestructural features ledtohigher

α

^and^RS valuesasdepictedinFig.5Band Fig.S4B.

Inaddition to thechromatograms foranalytes11-19,Figure 6 depictsselected chromatogramsforanalytes1-10 and12 aswell representingtheseparationsobtainedwithinthreeminutes.Using enantiopure analytes, elution sequences for analytes 2, 5, and 6 were determined and found to be the same forall columnsand mobilephases,theywere,R<S.

According to the data in Tables S1–S4, the separation factors, despite similar retention times and retention factors of the ﬁrst eluting enantiomers, sometimes differ considerably on the teicoplanin-andteicoplaninaglycone-basedCSPs,indicatingapos- sibledifferenceintheseparationmechanism.Inmostcases,higher selectivities and resolutions were obtained with the aglycone- based CSP underall the studied conditions,while no cleartrend could be observed for the variation in the retention times. As described earlier [35] the sugar units of the native teicoplanin may affectthe chiral recognition process in differentways; they block the possible interaction sites on the aglycone, occupy the space inside the “basket”, and offer additional interaction sites since the three sugar units are themselves chiral. To quanti- tatively determine the effects of the sugar units, the equation ⁽^G^°) = −RT ln

α

^was ^applied ^for ^the calculation of the differences in enantioselective free energies between the two CSPs [⁽^G^°)TAG−⁽^G^°)T].AsillustratedinFig.7,theenergydiffer- ences[⁽^G^°)TAG−⁽^G^°)T]withveryfew exceptions,areneg- ative,i.e., theinteractionbetweenthe freeaglyconebasket(with- out the sugar moieties) and analyte improves chiral recognition.

Bycomparing the[⁽^G^°)TAG − ⁽^G^°)T] valuesforanalytes1- 6, it is interesting to note that in the case of molecules with a largersize, interactionsbetweenselector andanalyteare favored.

It should be noted that [⁽^G^°)TAG − ⁽^G^°)T] valuescan vary withtheamountofmobilephaseadditives.

3.5. vanDeemteranalysis

Organiccomponentsofeluents(MeOHandMeCN)usedinthis studyincombinationwithwateryieldmobilephaseswithconsid- erableviscosity, while combinationofMeOH andMeCN resultin low-viscosityeluentallowinghigherﬂowrateswithouthighback- pressures.AccordingtoDarcy’slaw,backpressurerelatestomobile phase viscosity and linear velocity [21,36]. For the investigation of vanDeemter plots,mobile phasespossessing low andmoder- ate viscosity [mobile phase b, aq.TEAA/MeOH (30/70 v/v) and c, MeCN/MeOH(30/70v/v,respectively)bothcontaining2.5mMTEA and5.0mMAcOH] wereselected, andplots wereconstructed on allfourstudiedcolumnsforanalytescontaininganaliphatic(6)or

(6)

Table 1

Selected chromatographic data for the separation of ß²-amino acids

Analyte Column Mobile phase k 1 α ^RS

aliphaticß²-amino acids

1 T-3.0 H 2O/MeOH (30/70 v/v ) 0.40 1.30 1.01

T-3.0 H 2O/MeOH (10/90 v/v ) 0.20 1.46 0.90

TAG-3 H 2O/MeOH (30/70 v/v ) 0.70 1.30 1.05

TAG-3 H 2O/MeOH (10/90 v/v ) 1.82 1.22 0.72

TAG-3.0 aq. TEAA/MeOH (10/90 v/v ) 1.75 1.24 0.74

2 T-3.0 aq. TEAA/MeOH (30/70 v/v ) 0.53 1.35 1.32

TAG-3.0 H 2O/MeOH (30/70 v/v ) 0.50 1.54 1.35

3 T-3.0 H 2O/MeOH (30/70 v/v ) 0.34 1.75 2.48

T-2.1 H 2O/MeOH (30/70 v/v ) 0.86 1.30 1.58

T-3.0 H 2O/MeOH (30/70 v/v ) 0.47 1.80 2.66

TAG-3.0 MeCN/MeOH (30/70 v/v ) 1.49 1.63 1.71

TAG-2.1 MeCN/MeOH (30/70 v/v ) 1.04 1.72 1.53

4 T-3.0 H 2O/MeOH (30/70 v/v ) 0.33 1.76 2.07

TAG-3.0 MeCN/MeOH (30/70 v/v ) 1.37 1.60 1.70

TAG-2.1 MeCN/MeOH (30/70 v/v ) 0.97 1.66 1.40

5 T-3.0 H 2O/MeOH (30/70 v/v ) 0.30 1.85 1.71

TAG-3.0 H 2O/MeOH (30/70 v/v ) 0.45 1.90 1.35

TAG-3.0 MeCN/MeOH (30/70 v/v ) 1.02 1.73 1.37

TAG-3.0 MeCN/MeOH (10/90 v/v ) 1.09 1.51 1.58

6 T-3.0 H 2O/MeOH (30/70 v/v ) 0.28 2.03 2.50

T-3.0 MeCN/MeOH (10/90 v/v ) 1.45 1.45 0.83

T-2.1 H 2O/MeOH (30/70 v/v ) 0.85 1.46 1.64

TAG-3.0 H 2O/MeOH (30/70 v/v ) 0.95 2.06 2.39

TAG-3.0 MeCN/MeOH (30/70 v/v ) 1.15 1.82 2.02

TAG-2.1 H 2O/MeOH (30/70 v/v ) 0.77 1.48 1.72

TAG-2.1 MeCN/MeOH (30/70 v/v ) 0.69 1.90 1.89

7 T-3.0 H 2O/MeOH (30/70 v/v ) 0.27 1.81 1.97

TAG-3.0 H 2O/MeOH (30/70 v/v ) 0.66 1.64 1.80

TAG-3.0 MeCN/MeOH (30/70 v/v ) 1.08 1.60 1.77

( continued on next page )

(7)

Table 1 ( continued )

Analyte Column Mobile phase k 1 α ^RS

8 T-3.0 H 2O/MeOH (30/70 v/v ) 1.10 1.59 2.05

T-3.0 aq. TEAA/MeOH (10/90 v/v ) 0.79 1.61 1.67

TAG-3.0 MeCN/MeOH (30/70 v/v ) 1.55 1.56 1.75

aromatic ß²-amino acids

9 T-3.0 H 2O/MeOH (30/70 v/v ) 0.82 1.36 1.75

T-3.0 aq. TEAA/MeOH (30/70 v/v ) 0.62 1.48 1.99

T-2.1 H 2O/MeOH (30/70 v/v ) 1.73 1.23 1.94

T-2.1 aq. TEAA/MeOH (30/70 v/v ) 0.69 1.53 1.73

T-2.1 MeCN/MeOH (30/70 v/v ) 2.43 1.39 1.68

TAG-3.0 H 2O/MeOH (30/70 v/v ) 1.00 1.49 1.82

TAG-2.1 H 2O/MeOH (30/70 v/v ) 1.09 1.34 1.59

TAG-2.1 MeCN/MeOH (30/70 v/v ) 1.67 1.48 1.68

10 T-3.0 H 2O/MeOH (30/70 v/v ) 0.97 1.50 1.74

T-3.0 aq. TEAA/MeOH (30/70 v/v ) 0.62 1.70 2.77

T-3.0 MeCN/MeOH (30/70 v/v ) 1.88 1.43 1.76

T-2.1 H 2O/MeOH (30/70 v/v ) 1.74 1.21 1.87

T-2.1 aq. TEAA/MeOH (30/70 v/v ) 0.72 1.72 2.40

T-2.1 MeCN/MeOH (30/70 v/v ) 2.07 1.52 2.08

TAG-3.0 H 2O/MeOH (30/70 v/v ) 1.19 1.73 2.83

TAG-3.0 MeCN/MeOH (30/70 v/v ) 2.05 1.71 2.06

TAG-2.1 H 2O/MeOH (30/70 v/v ) 1.24 1.46 2.31

TAG-2.1 MeCN/MeOH (30/70 v/v ) 1.45 1.80 2.54

11 T-3.0 aq. TEAA/MeOH (30/70 v/v ) 0.64 1.39 1.70

T-2.1 H 2O/MeOH (30/70 v/v ) 1.87 1.17 1.58

T-2.1 aq. TEAA/MeOH (30/70 v/v ) 0.72 1.46 1.59

T-2.1 MeCN/MeOH (30/70 v/v ) 2.30 1.33 1.46

TAG-3.0 MeCN/MeOH (30/70 v/v ) 1.02 1.41 1.78

TAG-2.1 H 2O/MeOH (30/70 v/v ) 1.39 1.39 2.02

TAG-2.1 MeCN/MeOH (30/70 v/v ) 1.58 1.45 1.64

12 T-3.0 aq. TEAA/MeOH (30/70 v/v ) 1.57 1.17 1.27

T-3.0 aq. TEAA/MeOH (10/90 v/v ) 1.88 1.44 1.97

T-2.1 MeCN/MeOH (30/70 v/v ) 2.95 1.32 1.37

T-2.1 MeCN/MeOH (10/90 v/v ) 2.35 1.27 1.19

TAG-3.0 MeCN/MeOH (30/70 v/v ) 2.62 1.33 1.06

TAG-3.0 MeCN/MeOH (10/90 v/v ) 3.08 1.39 1.15

TAG-2.1 MeCN/MeOH (20/80 v/v ) 1.81 1.43 1.53

(8)

Analyte Column Mobile phase k 1 α R S

13 T-3.0 aq. TEAA/MeOH (30/70 v/v ) 0.73 1.18 1.32

T-3.0 aq. TEAA/MeOH (10/90 v/v ) 1.30 1.32 1.56

T-2.1 aq. TEAA/MeOH (30/70 v/v ) 0.78 1.30 1.32

T-2.1 aq. TEAA/MeOH (10/90 v/v ) 1.62 1.40 1.91

TAG-2.1 H 2O/MeOH (30/70 v/v ) 1.60 1.28 1.47

14 T-3.0 aq. TEAA/MeOH (30/70 v/v ) 0.67 1.46 1.84

T-2.1 H 2O/MeOH (30/70 v/v ) 1.71 1.16 1.49

T-2.1 aq. TEAA/MeOH (30/70 v/v ) 0.73 1.51 1.74

T-2.1 MeCN/MeOH (30/70 v/v ) 2.47 1.37 1.59

TAG-3.0 H 2O/MeOH (30/70 v/v ) 0.83 1.48 1.89

TAG-3.0 MeCN/MeOH (30/70 v/v ) 2.63 1.43 1.38

TAG-2.1 H 2O/MeOH (30/70 v/v ) 1.11 1.39 2.09

TAG-2.1 MeCN/MeOH (30/70 v/v ) 1.88 1.53 1.91

15 T-3.0 H 2O/MeOH (30/70 v/v ) 0.86 1.40 1.57

T-3.0 aq. TEAA/MeOH (30/70 v/v ) 0.59 1.50 3.10

T-3.0 MeCN/MeOH (30/70 v/v ) 2.31 1.36 1.38

TAG-3.0 H 2O/MeOH (30/70 v/v ) 0.82 2.10 2.58

16 T-3.0 aq. TEAA/MeOH (30/70 v/v ) 0.68 1.34 1.51

T-2.1 H 2O/MeOH (30/70 v/v ) 1.92 1.16 1.50

T-2.1 aq. TEAA/MeOH (30/70 v/v ) 0.76 1.41 1.48

TAG-3.0 H 2O/MeOH (30/70 v/v ) 1.12 1.37 1.48

TAG-2.1 H 2O/MeOH (30/70 v/v ) 1.47 1.34 1.90

TAG-2.1 MeCN/MeOH (30/70 v/v ) 1.63 1.41 1.51

17 T-3.0 aq. TEAA/MeOH (30/70 v/v ) 0.67 1.45 1.82

T-2.1 H 2O/MeOH (30/70 v/v ) 1.86 1.18 1.57

T-2.1 aq. TEAA/MeOH (30/70 v/v ) 0.73 1.50 1.74

TAG-3.0 H 2O/MeOH (30/70 v/v ) 0.92 2.00 2.91

TAG-3.0 MeCN/MeOH (30/70 v/v ) 2.91 1.70 2.14

TAG-2.1 H 2O/MeOH (30/70 v/v ) 1.26 1.74 3.00

TAG-2.1 MeCN/MeOH (30/70 v/v ) 2.13 1.83 2.64

(9)

Analyte Column Mobile phase k 1 α ^RS

18 T-3.0 H 2O/MeOH (30/70 v/v ) 1.19 1.35 1.58

T-3.0 aq. TEAA/MeOH (30/70 v/v ) 0.74 1.47 2.09

T-3.0 MeCN/MeOH (30/70 v/v ) 2.37 1.32 1.48

T-2.1 H 2O/MeOH (30/70 v/v ) 2.02 1.22 1.91

T-2.1 aq. TEAA/MeOH (30/70 v/v ) 0.84 1.52 1.94

T-2.1 MeCN/MeOH (30/70 v/v ) 2.71 1.40 1.74

TAG-3.0 H 2O/MeOH (30/70 v/v ) 1.05 1.73 2.72

TAG-3.0 aq. TEAA/MeOH (30/70 v/v ) 1.07 1.66 2.75

TAG-3.0 MeCN/MeOH (30/70 v/v ) 2.50 1.55 1.68

TAG-2.1 H 2O/MeOH (30/70 v/v ) 1.44 1.56 2.88

TAG-2.1 MeCN/MeOH (30/70 v/v ) 1.76 1.66 2.27

19 T-3.0 H 2O/MeOH (30/70 v/v ) 1.17 1.45 1.98

T-3.0 aq. TEAA/MeOH (30/70 v/v ) 0.75 1.64 2.71

T-3.0 MeCN/MeOH (30/70 v/v ) 2.27 1.36 1.55

T-2.1 H 2O/MeOH (30/70 v/v ) 1.97 1.31 2.37

T-2.1 aq. TEAA/MeOH (30/70 v/v ) 0.88 1.70 2.57

T-2.1 MeCN/MeOH (30/70 v/v ) 2.96 1.45 1.94

19 TAG-3.0 H 2O/MeOH (30/70 v/v ) 1.53 1.75 2.77

TAG-3.0 MeCN/MeOH (30/70 v/v ) 2.95 1.53 1.50

TAG-2.1 H 2O/MeOH (30/70 v/v ) 1.60 1.67 3.22

TAG-2.1 MeCN/MeOH (30/70 v/v ) 2.12 1.69 2.30

Chromatographic conditions: column, T-3.0, T-2.1, TAG-3.0 and TAG-2.1 ; mobile phase, H 2O/MeOH (30/70 v/v ), aq. TEAA/MeOH (30/70 v/v ) and MeCN/MeOH (30/70 v/v ), the latter two contain 2.5 mM TEA and 5.0 mM AcOH; ﬂow rate, 0.3 ml min ⁻¹; detection, 210-258 nm; temperature, 20

°C

aromatic(9)sidechain.vanDeemterplotsareshowninFigure8A (foranalyte6)andFig.S5A(foranalyte9)inpolar-ionicmode.In the polar-ionic mode, the curvesfor the ﬁrst elutingenantiomer showcharacteristicminimaforanalyte6onT-3.0,T-2.1,andTAG- 3.0 columns, and a slight minima on TAG-2.1 at ~1.5 mm sec^–1 (Fig. 8A). It should be notedthat 2.1 mm i.d. columns are usu- allylesseﬃcientthan3.0mmonesduetowall effects(Fig.8A).

The Hminima onT-3.0 andTAG-3.0 were registeredat0.24 mm sec^–1,whileonT-2.1at0.48mmsec^–1linearvelocity,whichcorre- spondstoaﬂowrateof0.1mlmin^–1.ThevanDeemtercurvesfor teicoplanin-basedcolumns runbelowthe plotsofthe teicoplanin aglycone. Fig. S5 A depicts van Deemter plots for analyte 9 un- derthesameconditions.Theshapeofthecurveforcolumnswith 3.0 mm i.d. are similar to plots obtained for analyte 6 (minima areintherange0.24–0.48mmsec^–1,i.e.,0.1–0.2mlmin^–1),while plots obtainedoncolumns with2.1mmi.d. exhibitedslightmin- imaatlowerﬂowrates(0.05–0.1mlmin^–1).Interestingly,theH-u plot for theteicoplanin aglycone column with2.1 mm i.d. (TAG- 2.1)runsbelowthesametype ofcolumnwithalarger i.d.(TAG- 3.0). Figures8B andS5B depict vanDeemter plotsforanalytes 6 and 9 applying mobile phase b, aq.TEAA/MeOH (30/70 v/v) containing 2.5 mM TEA and 5.0 mM AcOH on teicoplanin- and te-

icoplaninaglycone-basedcolumnspossessingdifferentinternaldi- ameters.ThevanDeemtercurvesathighflowrates(wheretheC- termdominates) on T-3.0 columns exhibited a slight increase in plate height, while on T-2.1 columns a decrease in plate height (slightly negativeslope) wasregistered for both analytes athigh flowrates.Itisdescribedseveraltimesthatathighbackpressures, two typesoftemperature gradients – axialandradial – exist to- getherastheresultofsignificantfrictionalheating[16,37-39].Ax- ialtemperaturedifferencesrangingfrom11to16°Ccanreadilybe generatedwhenpressureabove300barisapplied[16,37].Insome cases,longitudinalfrictionalheatingwasfoundtoincreasethechi- ral resolution when small particles andhigh flow rates are used [16,21].InFig.8C,vanDeemterplotsforthefirstandsecondelut- ingenantiomer ofanalyte6on TAG-3andanalyte9on theT-2.1 columnaredepicted.Itisinteresting tonote thatidenticalkinetic plotshapeswererecordedforbothenantiomerswiththecurvefor the second enantiomershifted upwards.The similar shapesindi- catethat bothenantiomershavesimilar adsorption/desorptionki- netics(thesameresultswereobtainedunderotherconditionstoo;

datanotshown).Insummary,comparingresultsobtainedforvan Deemteranalysesandscreeningexperimentsof19ß²-aminoacids (registeredata ﬂow rateof0.3 mlmin^–1), thefollowing conclu-

(10)

Figure 5. Effect of nature of substituents and chemical structure of analytes on chromatographic performance for analytes 11 and 13-19 Chromatographic condition, column, TAG-3.0; mobile phase, A , H 2O/MeOH (30/70 v/v ), B, aq. TEAA/MeOH (30/70 v/v ), concentration of TEA and AcOH in mobile phase 2.5 and 5.0 mM, respectively and the actual pH of the mobile phase, pH a5.5; ﬂow rate, 0.3 ml min ^–1; detection, 258 nm; temperature, 20 °C

Figure 6. Selected chromatograms for analytes 1-10 and 12 Chromatographic conditions, columns, for analytes 1, 4, 6, 7 , and 8 T-3.0, for 2, 3 and 5 TAG-3.0, for 9 and 10 T-2.1 and for 11 and 12 TAG-2.1; mobile phase, for analytes 1-7 and 11 , H 2O/MeOH (30/70 v/v ), for 8-10 and 12 aq. TEAA/MeOH (30/70 v/v ), concentration of TEA and AcOH in mobile phase 2.5 and 5.0 mM, respectively and the actual pH of the mobile phase, pH a5.5; ﬂow rate, 0.3 ml min ^–1; detection, 258 nm; temperature, 20 °C;

sions can be drawn: (i) higher plate numbers were obtainedon teicoplanin-based than on teicoplanin aglycone-based CSP (T-3.0 vs. TAG-3.0 andT-2.1 vs.TAG-2.1), (ii) in general, forSPPs of 2.7 μm, thenarrow borecolumns (2.1 mm i.d.)show decreased eﬃ- ciencycomparedtotheircounterpartswith3.0mmi.d.Note,that the lattercolumns were expected to outperform the columns of 2.1 mm i.d, and this expectation was met underall the studied conditions. It mustbe emphasized, however, that columnperfor- mance,inthepractice,dependsonboththenatureofanalytesand themobilephasecomposition.Hvaluesforanalytespossessingan alkyl sidechain(1-8) werealwayssmalleroncolumnsof3.0mm i.d., while for analytespossessing an aromatic side chain (9-19), columns of 2.1mm i.d. showed better performance (Table S1–S4 andFig.S5A).However,intheRPmodeforanalytes9-19,columns of 3.0 mm i.d. always outperformed the columns of 2.1 mm i.d.

columns(TableS1–S4).

3.6. Temperaturedependenceandthermodynamicparameters

Studyingtheeffectsoftemperatureonretentionandenantios- electivity in chiral separations is an often applied methodology to gather information on enantiomer recognition [40-43]. Theo- retically,retention observed on chiral CSPsconsists ofchiral and nonchiral components [44-48],however, in thisstudy, these two components are not differentiated. Keeping in mind the limita- tionsoftheapproach,usedherein thedifferenceinthechangein standardenthalpy⁽^H^°)ândêntropy⁽^S^°)^for^theênantiomer pairswere calculatedusingtherelationshipbetweenln

α

^(natural

logarithmoftheapparentselectivityfactor)andT⁻¹(reciprocalof absolutetemperature)asdescribedbythevan’tHoff equation:

ln

α

⁼ ⁻

⁽

_RT

^H^◦

⁾

⁺

⁽

_R^S^◦

⁾

⁽²⁾

(11)

Figure 7. Enantioselectivity free energy differences ( G °) TAG −( G °) Tbetween aglycone and native teicoplanin selector Chromatographic condition, column, A , TAG-3.0 vs. T-3.0 and B, TAG-2.1 vs. T-2.1; mobile phase, a , H 2O/MeOH 30/70 v/v ), b , aq. TEAA/MeOH (30/70 v/v ), concentration of TEA and AcOH in the mobile phase 2.5 and 5.0 mM, respectively, and the actual pH of the mobile phase, pH a5.5, c , MeCN/MeOH (30/70 v/v ), concentration of TEA and AcOH in the mobile phase 2.5 and 5.0 mM, respectively;

ﬂow rate, 0.3 ml min ^–1; detection, 210-258 nm; temperature, 20 °C; symbols, mobile phase, a , mobile phase, b and mobile phase, c

Figure 8. Plots of plate heights versus superficial velocities for analytes 6 and 9 on macrocyclic glycopeptide-based columns Chromatographic conditions: columns, A , T-3.0, T-2.1, TAG-3.0 and TAG-2.1, B , T-3.0, T-2.1 and C , TAG-3.0 and T-2.1 ; mobile phase, A , MeCN/MeOH (30/70 v/v ), concentration of TEA and AcOH in the mobile phase 2.5 and 5.0 mM, respectively, B and C , aq. TEAA/MeOH (30/70 v/v ), concentration of TEA and AcOH in the mobile phase 2.5 and 5.0 mM, respectively, and the actual pH aof the mobile phase, pH a5.5; detection, 210-258 nm; temperature, 20 °C; symbols, A , analyte 6 , T-3.0, T-2.1, TAG-3.0, TAG-2.1; B , analyte 6 , T-3.0, T-2.1, analyte 9 , T-3.0, T-2.1; C , analyte 6 , TAG-3.0 (first enantiomer), TAG-3.0 (second enantiomer), analyte 9 , █T-2.1 (first enantiomer), T-2.1 (second enantiomer)

where Ristheuniversal gasconstant. Asdiscussed earlier,under UHPLC conditionsoperating withinlet pressures above 300 bars, the generatedaxialtemperaturedifferencescanlead toamarked difference betweenreal operationaland setconditions [16,21,37].

Toavoidthetemperaturedifferencescausedbyhighbackpressure, thenatureofappliedmobilephasesandﬂowrateswere carefully selected forthis study.For example,CSPswith 2.1mm i.d. were appliedonlyinpolar-ionicmode.

Dependence of the chromatographic parameters on temperature was studied on the four columns for analyte 8 possessing an aliphatic side chain, and for analyte 9 bearing an aromatic

side chain in the temperature range 5-50 °C. Experimental data withmobilephasesofaq.TEAA/MeOH(30/70v/v)andMeCN/MeOH (30/70v/v)bothcontaining20mMTEAAarepresentedinTableS5.

Asmostfrequentlyobserved,bothkand

α

^decreased^with^increas-

ingtemperatureinallcases.Resolution usuallydecreaseswithin- creasingtemperature, whilein a few cases, R_S,exhibited a max- imumcurve withthe changeoftemperature(Table S5). LowerR_S valuesatlowandhightemperaturescanbeattributedtothelower kineticandhigherthermodynamiceffect(decreased

α

^values),^re-

spectively.